Sweetgreen's Comps Flipped From Mix to Traffic in One Quarter — and 500 Basis Points of That Drop Was a Lettuce Scare

June's -6.2% and September's -6.4% look like the same bad quarter printed twice. The build was different. Traffic deterioration of 4.9% was 2.9 points worse than June's 2.0%, but product-mix damage improved by 2.1 points and price added 0.6 points for the first time in two quarters. The headline barely moved while its composition flipped. That distinction matters for how the trade-down story gets read: mix is a price-value signal that management can engineer, while July's traffic was an outside event that arrived from a farm in central Mexico.

On July 18, 2026 the Food and Drug Administration named iceberg lettuce processed at Taylor Farms' central Mexico facility as the likely source of a cyclospora outbreak that sickened at least 10,000 people, and Taylor Farms recalled product. Sweetgreen's own food was not implicated; management described the chain as caught by association. The damage ran through the whole salad category. Consumer concern cost Sweetgreen roughly six percentage points of July same-store sales, and the fear spread well beyond one chain: Chopt traffic fell 24% on July 18 against its day-of-week average, and NielsenIQ measured dollar sales of prepackaged salad down 14% in the four weeks ended July 25.

The most informative disclosure in the quarter is not in the press release. Owned digital revenue rose about 10% year over year to $66.9 million and marketplace revenue rose about 8% to $49.5 million, while in-store revenue fell about 9.5% to $59.6 million. Total revenue still grew 2.1% only because of those two digital lines. The traffic loss is a physical-visit problem, not a demand problem: a delivery order never asks the guest to inspect a head of lettuce, which is exactly what the July headlines were about.

Units are where the real damage sits. Sweetgreen opened three net new restaurants in the quarter against six a year earlier, and the comparable base excluded eleven restaurants with closures in the period, nine of them permanent, so the -6.4% describes only the survivors of a shrinking estate. Restaurant-level profit margin fell 270 basis points to 10.4% on lower sales and higher labor costs, and adjusted EBITDA widened to -$6.8 million from a -3.8% margin. Management is buying traffic back with price points while the unit engine stalls, and a comp that has to be manufactured by discounting is worth less than one that compounds.

Wonder is the supply-chain relationship that matters most for the equity, and it cuts both ways. In the first quarter of 2026 Sweetgreen sold Spyce and the Infinite Kitchen technology to Wonder for $186.4 million, booked a $160.6 million pre-tax gain, and now holds $86.4 million of Wonder Series C preferred stock while depending on Wonder to supply the kitchen units behind roughly half of this year's openings. The company traded ownership of its automation for cash and a vendor relationship it no longer controls, and the equity it received is private and illiquid.

Net loss narrowing from $36.1 million to $30.7 million reads as progress, but the only profitable stretch on the trailing books is a byproduct of selling the robotics division. Trailing net income of $19.1 million includes the $160.6 million pre-tax gain on the Spyce disposal, and operating income over the same twelve months was -$132.5 million. Cash was $119.8 million at Sept. 27, 2026 against $350.0 million of finance lease obligations, net debt of $230.3 million, and trailing operating cash outflow of $37.1 million.

Over the next two quarters the swing factor is arithmetic: a full-year guide of about -5.5% against a -6.4% third quarter requires roughly +3.4% in the December quarter, which means October's +5% has to survive two more months and a harder seasonal comparison.

Over one to three years the question is whether roughly ten openings a year, then about five, can ever be replaced by same-store sales, and whether the mix and margin repair holds once the shock comps are gone.

For fast casual as a group, the Sweetgreen print is good news and bad news at once. Good: the category is not in a consumer retreat. The National Restaurant Association found 54% of operators reported same-store sales gains between September 2025 and September 2026, and Chipotle Mexican Grill and CAVA Group both grew traffic in the very quarter Sweetgreen lost it. Bad: the traffic that left Sweetgreen went to chains with lighter delivery mix and, in CAVA's case, a model that also sells bowls through grocery retail. The trade-down is a mix problem inside one chain, not a demand problem across the category.